Budgeting

An account imported twice: compare options on equal terms

Compare overlapping exports on account coverage, identifiers, dates and reconciliation evidence before choosing a dependable working source.

Editorial illustration for An account imported twice: compare options on equal terms

Compare the information, not the filename

When the same account has been imported twice, choosing a working source requires more than keeping whichever file was downloaded last. First identify the purpose of the comparison: finding overlapping events, recovering missing references or establishing a complete interval. A newer file may cover less history, while an older file may preserve a useful detail that a different export omits.

List the account, opening and closing dates, transaction count and any stated balance information for each source. Keep the original files unchanged. The comparison should explain what each file contributes before any rows are removed from the working record. File format alone does not establish completeness or accuracy.

Put fields on comparable terms

Map the fields that describe account identity, debit or credit direction, amount, dates and references. Document whether a date means purchase, posting or export. A comparison between different date meanings can produce apparent gaps that are actually differences in presentation.

Preserve raw values alongside any normalised comparison fields. For example, trimming spaces from a reference may help locate a match, while removing a leading zero could change its meaning. Do not alter the original source to make a matching method easier. A useful normalisation can always be explained and traced back to the value received.

Measure overlap without equating row counts

Suppose, illustratively, export A contains 40 rows for a period and export B contains 35. Investigation establishes that 30 rows describe the same events across both files, with ten events appearing only in A and five only in B. The combined unique-event count is 45, not 75 and not automatically 40.

This calculation assumes all 30 overlap decisions are supported and every source row represents an individual event under the comparison's definition. If one export groups payments into summaries, row counts are no longer directly comparable. Identify the grouped representation before attempting to use count differences as evidence of missing or duplicated transactions.

Evaluate the quality of identifiers

A source containing stable event references can be easier to reconcile than one containing only dates and merchant descriptions. However, a reference should be checked within its account and source context. Do not assume that matching text has the same meaning across unrelated exports or accounts.

Assess how unresolved pairs will be handled before choosing a working source. A method that confidently deletes all repeated amounts may appear efficient while erasing genuine purchases. Prefer a comparison process that makes ambiguity visible and preserves supporting evidence. The best source for one investigation may still need information from the other file to resolve a specific event.

Check coverage against the question

If the task is a full-month report, confirm that the selected working set covers the entire month according to the relevant date meaning. A file ending before the last posting day cannot establish a complete month merely because its total resembles the previous month. Keep any uncovered interval visible.

Review credits and transfers as well as purchases. Two files can have matching spending subtotals while differing materially on incoming money or account movements. Where balance evidence exists, reconcile the complete event set appropriate to that evidence. Do not use a narrow purchase comparison to claim that the whole account has been fully reconciled.

Record why the working set was selected

Describe which source supplies the main event list, which supplementary fields were retained and how overlapping rows are connected. Keep a reference to both original exports. Selecting one working representation should not require discarding the other source's useful evidence or hiding an unresolved discrepancy.

Before the next download, repeat the comparison on the new interval rather than assuming the previous selection remains complete forever. Export coverage and presentation can differ between downloads. A dependable choice is one supported by the current files and the reporting question, with a reproducible explanation of overlap, omissions and remaining uncertainty.

SPENDIXAI JOURNAL

Useful ideas start with a clearer picture of your own spending.

Explore your reports